362 THE CANADIAN ENTOMOLOGIST. 



5. Labium composed of two segments at most. The forms swim on their 



bellies. Ova laid externally 25. Corixidce. 



5a. Labium free, composed of three or four segments. Forms swim on 

 their backs. Ova inserted in stems, etc., of water- 

 plants 26. Notonectidce. 



Considering, then, the nymphal characters, the number of segments 

 in the tarsi, the condition of the tegmina, and of the stink-glands, etc., it 

 seems to me that the only families that have any claims to be considered 

 typical are the Cimicidae, Pyrrhocoridas, Myodochidae, Nabidae and 

 Anthocoridae. 



The Anthocoridae are undoubtedly the most generalized of the 

 Pagiopoda, and are probably very ancient,* but I think that the Pagiopoda 

 are less typical, as a whole, than the Trochalopoda. The Nabidae have 

 less highly-developed stink-glands, and, I think, are a development of 

 some "Protomyodochidae" in the direction of greater rapacity and agility. 

 The ocellated ancestors of the Pyrrhocoridae might almost as well, how- 

 ever, as the Cimicidae, ■(' be the most typical Heteroptera. It is the fact 

 that certain Cimicid nymphs have the first orifice paired, that leads me to 

 propose that place for them, and it will scarcely be questioned that it is 

 much more likely that a paired gland should become coalesced, than that a 

 single gland should be divided in such a case. 



The Heteroptera are admittedly one of the most ancient and isolated 

 groups of insects, but they were probably much more dominant in past 

 times than to-day. I believe that the families which are usually regarded 

 as the "highest" are really so, but are also at the same time nearest to the 

 typical Heteropteron of the later Paleozoic times, when it had finally 

 separated itself from the Homoptera, and a long time since from the other 

 Insecta. 



The subjoined table shows, in some sort, my views on the probable 

 phylogemy of the Heteroptera, but a good deal of latitude must be given 

 to the detailed lines of descent. 



The following shows my opinion as to the general relationships of the 

 26 families : 



*Reuter says : "J'ai commence mon ouvrag-e par la famille des Capsides 

 [Miridae], parceque celle-ci me parait etre la plus basse dans le syst^me " (Act. 

 Soc. Sci. Fenn., XIII, 6), but this was in 1878. See the P.S. at the end of this 

 paper. 



tin selecting- "Cimicidae," I refer specially only to the Cimicinas ( =- Asopinse). 

 The other subfamilies are mostly much more recently developed. 



